Welcome

In theory, I find Representation Learning and Online Learning fascinating.

In practice, only Video Understanding really captures my attention.

I am willing to converse and study with you.

But please, no silly questions.

Logic only.

See my publications on Google Scholar .

Pin2Top 🤩

We’re looking for interns in reinforcement learning, multimodal learning, and video understanding. Talented folks, come join us! If you are interested, please send your CV to zqsiat@gmail.com. What we offer: 1) Mentorship from experienced researchers; 2) Abundant computing resources; 3) Room to explore freely and innovate boldly.
History Intern:
1) Yusong Wang (friend and collaborator), PhD from Tokyo Tech, homepage.
2) Jiahao Nie (collaborator), PhD from NTU, homepage.
3) Yu Wei (collaborator), PhD from NTU, homepage.

News 🥳

• [2026/06] 1 Papers on Representation Learning “When Manifold-Constrained Hyper-Connections Meet Diffusion Transformers” is submited. 🎉
• [2026/05] 1 Papers on Segmentation “Cross-Domain Few-Shot Segmentation via Conditional Point Sparsification in SAM” is submited. 🎉
• [2026/05] 1 Papers on Video Understanding “Unlocking High-Frequency Temporal Visual Cues for Personality Traits Recognition” has been accepted for publication in IEEE Access.🎉
• [2026/05] 1 Papers on Representation Learning “LAViT: Injecting Locality into Vision Transformers for Robust Representation Learning” is submited.
• [2026/02] It’s a great honor for us: one paper accepted to CVPR main, and one to CVPR Findings. 🎉
• [2026/02] I commenced my postdoctoral research at the Computer Science and Technology Research Station of Fudan University under the guidance of Professor Zhang Junping.
• [2026/01] We are honored to have two papers accepted at ICLR 2026! 🎉
• [2026/01] 1 Papers on Multi-Modal “On the Consistency of 3D Large Language Models” is submited. 🎉
• [2025/10] Thrilled to announce I’ve joined TeleAI as an Algorithm Researcher, where I’ll be focusing on AI Flow (formerly referred to as 智传网).
• [2025/10] 1 Papers on AI4SCI (ExpVid: A Benchmark for Experiment Video Understanding & Reasoning) are submited.
• [2025/09] 1 Papers on Deep Learning (Person Identify Shift for Privacy-Preserving Person Re-identification) are accepted by SCIENCE CHINA Information Sciences, SCIS CCF-A.
• [2025/08] 1 Papers on Machine Learning (DPL++: Advancing the Network Performance via Image and Label Perturbations) are accepted by Transactions on Pattern Analysis and Machine Intelligence, TPAMI 2025.
• [2024/10] 1 Papers on Video Understanding (Learning Discriminative Representations in Videos) are accepted by IEEE Signal Processing Letters, SPL 2024.
• [2024/10] I have been invited to serve as a Reviewer for CVPR, CVPR 2025.
• [2024/09] 1 Papers on Video Understanding (Open-Vocabulary Online Action Detection) are accepted by Advances in Neural Information Processing Systems, NeurIPS 2024.
• [2024/08] I have been invited to serve as a Reviewer for AAAI, AAAI 2025.
• [2024/05] 1 Paper on visual classification (Code Needs Comments) is accepted by International Journal of Computer Vision, IJCV 2024.

Publications🗞️

Beyond Static Artifacts: A Forensic Benchmark for Video Deepfake Reasoning in Vision Language Models
Zheyuan Gu*, Qingsong Zhao*, Yusong Wang, Zhaohong Huang, Xinqi Li, Cheng Yuan, Jiawei Shao, Chi Zhang, Xuelong Li
CVPR 2026 Findings

Streaming Video Crime Anticipation with Spatio-Temporal Causal Reasoning
Yusong Wang, Zheyuan Gu, MAO KEYU, Minghao Shao, Mingkun Xu, Prayag Tiwari, Jiawei Shao, Qingsong Zhao†
CVPR 2026

RIVER: Real-time Video Interaction Benchmark
Yansong Shi*, Qingsong Zhao*, Tianxiang Jiang, Xiangyu Zeng, Yi Wang, Limin Wang
ICLR 2026

ExpVid: A Benchmark for Experiment Video Understanding & Reasoning
Yicheng Xu, Yue Wu, Jiashuo Yu, Ziang Yan, Tianxiang Jiang, Yinan He, Qingsong Zhao, Kai Chen, Yu Qiao, Limin Wang, Manabu Okumura, Yi Wang
ICLR 2026

Does Video-Text Pretraining Help Open-Vocabulary Online Action Detection?
Qingsong Zhao, Yi Wang, Jilan Xu, Yinan He, Zifan Song, Limin Wang, Yu Qiao, Cairong Zhao
NeurIPS 2024
[Code]

Adaptive Discriminative Regularization for Visual Classification
Qingsong Zhao, Yi Wang, Cairong Zhao
IJCV 2024
[Paper]

Learning Discriminative Representations in Videos via Active Embedding Distance Correlation
Qingsong Zhao, Yi Wang, Cairong Zhao
SPL 2024
[Code]

Unleashing the Potential of Mamba: Boosting a LiDAR 3D Sparse Detector by Using Cross-Model Knowledge Distillation
Rui Yu, Runkai Zhao, Jiagen Li, Qingsong Zhao
arXiv 2025
[Paper]

Exploiting Spatio-Temporal Representation for 3D Human Action Recognition from Depth Map Sequences
Xiaopeng Ji, Qingsong Zhao, Jun Cheng, Chenfei Ma
KBS 2021
[Paper]

A Cuboid CNN Model with an Attention Mechanism for Skeleton-based Action Recognition
Kaijun Zhu, Ruxin Wang, Qingsong Zhao, Jun Cheng, Dapeng Tao
TMM 2020
[Paper]

Chaos

数字图像处理

2018年04月10日 数字图像处理第三章作业 图像增强 LoG [code]
2019年07月30日 双线性插值原理 图像旋转
2019年07月31日 feature map以及反卷积的大小计算
2019年07月31日 中值滤波原理及其C++代码
2019年08月02日 加权中值滤波原理及其公式推导

机器学习

2018年07月15日 生物医学信息 心电数据分类识别. 2021年02月02日 二维卷积的python实现
2021年02月02日 softmax激活函数结合cross entropy损失函数的反向传播过程计算
2021年02月03日BP神经网络的反向传播实现 MNIST